Bessler Collins Gravity Wheel

[John Collins]

Two more pics. The lower one shows the red lever pushing the blue lever to the right, anti-clockwise.

The upper picture shows the same red lever lifting the blue lever in the leading or preceding segment.

JC

[John Collins]

I’ll post the connections and pulleys with cords later.

One more thing to note. The cords are always taught, in wubbly’s sim they go slack.

JC

[Wubbly]

Yes, Please post the pulley connections. I see the rope in Bessler's drawing and I don't know how you are interpreting that picture. Currently that picture is still saying nothing to me.

I can't see how there can be no slack when the red mass resets. If there is no slack, the red and blue masses would have to reset at the same time. Your pulley interpretation would help me understand.

Red lever extension too long? The script calculates the length of the short stub on the red lever (my R4). The red lever swings through a 90 degree arc and the blue lever swings through a 30 degree arc. If the red lever is going to lift the blue lever completely, the arc lengths must be equal. Instead of calculating equal arc lengths, I calculated the distance between the two endpoints of each mass to be equal as shown in my red lever drawing on page 2. (R1 = length of main red lever, R2 = length of the blue lever, R4 = 90 degree red lever extension). I can't model a round pulley. Unfortunately they are modeled as frictionless points.

The masses are equal, but one red mass drops through an arc of 90 degrees, and two blue masses raise through an arc of 30 degrees. Even though the lever lengths are different (1:1.6 ratio), do you see a problem with the potential energy lost vs the potential energy gain? For 1 kg masses and lever lengths of 1m and 1.6m I calculate an rough estimated energy loss of about 2 Joules when the red mass drops and the blue masses rise. (very rough estimate).

Do you know what angular velocity you are targeting for your wheel? In the following video you can see the latency affect. The red lever does not drop quite as shown in your pictures. There is a delay in the drop as the wheel rotates. See attached video for latency affect. Tracking was turned on so you can see this. The green lever is the "blue" lever in the current sector. If the red lever also pushed on the green lever to reset it, it would delay the movement of the red lever even more. At -100 deg/sec the red lever never fully extends as shown in the video.

As a comparison, Bessler's Gera and Draschwitz wheels ran at 50 RPM or more. This is an angular velocity of 300 degrees/second, and the sim only takes it up to one third of their angular velocities.

Latency affect of red lever drop Youtube video here: https://youtu.be/27QOkJq-cvg

[John Collins]

First of all can I thank you wubbly for the work you have done. I really appreciate the effort.

The picture shows a rope passing behind the wheel up to a pulley. It goes round the pulley and goes straight down to the chest of stones. The other end of the rope is shown wrapped around the axle of the other view of the wheel. It passes downwards at a slight angle to the other pulley. It goes round it and heads upwards at an angle behind the wheel. It is that exact shape that Bessler used between the red lever stub, and the blue lever in the leading mechanism.

The reason why the red lever can reset without affecting any other levers is be cause it moves back independently of any other lever.

The red lever stub has an arc which is close the neighbouring mechanism, so needs to be limited in length.

The cord which lifts the blue lever in the leading mechanism is attached at a point inwards from the weight itself, which allows the full 30 degree lift. It is a calculation best made on the actual build, as I found out.

The blue lever in the six o’clock mechanism isn’t lifted, it moves more or less horizontally just the same 30 degrees that it’s corresponding blue lever is lifted vertical from a position which equates the 8.30 on a clock face to 9.30, 30 degrees.

I cannot tell you anything about what angular velocity I’m expecting, I don’t know and I doubt I could work it out.

I know how difficult this is to get right, but instead of trying to work it out mathematically, not my strong point, I simply take the clues Bessler left for us and this design is what I was led to. Obviously I could have erred here or there or everywhere, I felt that the action I interpreted indicated a potential break in the symmetry, something we have all searched for, for ever!

JC

[Wubbly]

So you are looking at the angle of the rope in a picture and interpret that as angles between levers in mechanisms. OK.

Red lever Segment 1 to blue lever segment 1

As the red lever falls, the Red lever in segment 1 pushes on the Blue lever in segment 1 to reset the blue lever counter clockwise 30 degrees. This probably happens around the 6 o'clock position.

In my sim I was lifting the blue lever in the current segment via a pulley system instead of some mechanical arrangement that you propose.

Red lever Segment 1 to blue lever segment 2

When the (red) initiator lever falls, it pulls a cord which is attached to the short leg. This cord runs around two pulleys and its other end is attached near the end of the (blue) secondary lever in the preceding mechanism. The (red) initiator lever lifts the (blue) secondary lever in the preceding mechanism 30 degrees by pulling on the cord.

The cord which lifts the blue lever in the leading mechanism is attached at a point inwards from the weight itself, which allows the full 30 degree lift. It is a calculation best made on the actual build, as I found out.

So the cord goes from the red stub lever, around two pulleys and to the preceeding blue lever.

But instead of connecting the cord right at the blue weight, connect it at a different point on the blue lever and make the red stub lever shorter. Probably wouldn't make a difference in the sim. The sim allows you to make bodies not collide so the red stub lever can be as long as needed.

The reason why the red lever can reset without affecting any other levers is because it moves back independently of any other lever.

One more thing to note. The cords are always taught

If the red lever stub is connected to the blue lever somewhere, the cord will go slack when the red lever resets. Or what am I still missing here?

-------------------------------

Not trying to crush your hopes but ...

If you build the sim with only red levers (which the sim allows you to do) and input the radius of a Gera wheel (0.7 meters) and slowly crank the angular velocity up, at around 180-190 degrees/second (30-32 RPM), the red levers become pinned to the outside of the wheel and don't fall back inward to reset. The Gera wheel ran at 300 degrees/second (50 RPM).

If you repeat the process for a Draschwitz wheel with a radius of 1.4 meters, the red weights pin to the outside somewhere in the 130 to 140 degree/second range (21-23 RPM). The Draschwitz wheel ran at over 40 RPM.

If you repeat the process for the Merseburg and Kassel wheels, each with a radius of 1.85 meters, the weights pin to the outside of the wheel somewhere around 110-120 degrees/second (18-20 RPM). The Merseburg wheel ran at 40 RPM and the Kassel wheel ran at 20-26 RPM.

Observation: If this red lever was in one of Bessler's wheels running at their respective angular velocities, it would be pinned to the outside of the wheel and would not fall inward to reset.

[Tarsier79]

I'll make some observations.

Having an L shape lever doesn't give any advantage you couldn't get with a pulley or some other mechanism. I see you are trying to redirect force 90 degrees. Sometimes simple is better, so It is not a negative, but will not be a requirement for a working wheel.

Also, I don't see anything special or new about this design. It looks like any other early design for OB on this forum.. "Height for width" etc. The red weight path looks like it is taking a lot from rotation.

A positive, I do like the look of the movement of the blue weighted levers in Wubbly's animation. With the levers swapping positions it looks like they are always seekig to have 2 balanced pairs plus a leftover lever (even if it seems to be on the wrong side for OB.) I am not sure how you intend to harvest that positive and turn it into rotation.

[justsomeone]

Two big issues with this design John.

First, on the descending side...

The red lever will easily shift the blue lever into its position but the shift won't take place until the blue weight is in the 6 o'clock position. With a clockwise rotation, by the time the shift takes place and is finished , the blue weight will still be in the 6 o'clock position. There will not be any gain in excess weight on the descending side.

Secondly on the ascending side...

The red weight will never quickly raise the blue weight! The red lever will gradually gravitate to the center while trying to lift the blue lever. A hanging weight, gravitating to the center will not lift a weight like the same falling weight. This will prevent the gain you are trying to achieve.

Ok, so there is a third issue ( with me anyway )

In my opinion, there isn't any clues in the wheel drawings and there isn't any clues in the portraits. Sorry John, just my opinion.

I still have the utmost respect for you and your Bessler research! Thanks for sharing and may God bless you and your family.

[John Collins]

OK, thanks for the comments. There’s a lot for me to think about. I guess one key issue is the pinning of the red levers to the edge of the wheel. Also the blue levers not moving past the six o’clock radius despite being pushed backwards because the wheel is moving forwards at the same time.

This design has always seemed to me more complex than Karl intimated, perhaps it needs to be simplified in some way

I’ll comment again after I’ve has a good think. I’m down but not out!

JC

[John Collins]

Quick thought. The red levers will become pinned the edge of the wheel once in rotation, so my idea that they had to be able reset by returning gradually as the wheel rotated was wrong.

So if this problem is to be overcome then the falling of the six o’clock red lever not only pushes its accompanying blue lever backwards, lifts the blue lever vertically in the segment in front...and resets its accompanying red lever. This simplifies the connection between the red lever and the blue lever at six o’clock radius.

Sounds impossible but the first blue lever moves horizontally, so would need less energy to move it just 30 degrees. The blue lever in the leading segment needs a swift lift, again just 30 degrees and its own red lever would not need lifting, but pulling back against CF.

JC

[Wubbly]

@Tarsier, It reminded me of Walter's napkin sketch HERE , but Walter never completely described the driver/horse mass which would equate to John's red lever.

[Fletcher]

ETA : I see the lads have posted while I was musing ;7)

Hi John .. my 2 cents.

I'm looking at your design from the perspective of a sim builder first and foremost. In doing that I tend to see pictures in my head of how the sim should work in 2D. 3D if I have too tho not necessary for your design.

Part of that 'processing' is to break the problems down into groups that need solving one at a time. I see your design as a game of two halves.

Ultimately we want asymmetric torque conditions to push the wheel thru at least 72 degrees of rotation and accelerate, gain momentum etc.

First see if this makes sense, and correct me if I am wrong.

The blue lever-weight (lw) loads are pivoted at the wheel CoR. That means when they can be in freewheeling mode (acting like a pendulum staying vertical beneath a pivot) they do not contribute to system torque i.e. they are hanging beneath the axle freewheeling CCW.

The red lw drivers are pivoted at a radius from the CoR. They fall under gravity when they have reached their tipping point and can fall a vertical distance to stops. When they fall they do two things ..

1. lose vertical height (GPE).
2. depending on where they reach their stop and arc covered they can be in a position to create some torque. A small amount or large amount depending on angles and height lost within their sector. More vertical height lost and less torque contribution. More horizontal shift and more torque contribution but less work capability.

On their own it seems unlikely that the red lw driver movement alone will create the asymmetric torque desired - stating the obvious.

So you have included the second lever system of blue lws.

In the current sector (6 o'cl) the blue lw is desired to move backwards relative to the wheel body. This is to allow for it to be lifted at around 8.30 o'cl by the red driver in the lower sector.

Options : the current sector blue load can be actively pulled or pushed backwards or it can be left to freewheel backwards into position. But the latter option has no positive affect on system CoG/CoM. The former actively pulling it via a pulley system or combination SB-pulley system may help the system torque. If pulling it away from vertical it will have some GPE raising requirement which costs some works in Joules, even if small.

Next we want to lift the reset blue load into a new position at around 8.30 o'cl. This is almost a straight vertical lift. That requires Joules of work. It can be done provided the GPE lost by the driver is in excess of the lift GPE gain required for the blue load.

So we have a trade-off situation between driver height lost and its inherent torque contribution after shift, to height of lift required for blue load at 8.30 o'cl. Bearing in mind that the almost vertical lift in the leading sector gives almost no torque assistance to the system GoG/CoM at that stage.

As an added layer to the cake, we have CP forces to consider. Rotate too fast and we may have a negative affect on resetting the red driver at a good time and it in-fact may contribute negatively to system CoG/CoM overall.

So in my estimation there are 3 main things to think about for a design incorporating these elements.

A. Height lost and gained from lever shifts, and when - work capability.
B. Torque contributions from shifts - aim is for asymmetric torque to gain wheel momentum.
C. Centrifical Forces - possibly degrading or aiding performance.

That means there are limited permutations of connectivity options to explore.

I suggest let the current sector blue load freewheel backwards first, and concentrate on solving the connectivity and timing issues, and CF's issues, with the leading sectors.

If successful (or not) and want more torque contribution (system CoM) see if it is possible to actively pull the lower sector blue load back for reset secondly.

At least that is how I would attack things with multiple sim variations to attempt some progress towards a clear goal.

[John Collins]

Thank you for your concise and accurate account Fletch. Very useful as was wubbly’s sim showing the red levers pinned against the rim by cf.

The cf effect has to be overcome or negated. The old image of an ice skater comes to mind. Out stretched arms and spinning slowly, but when brought in, against cf, spinning speeds up.

I always knew that my interpretations of the clues might be wrong but I believed and still do, that the elements I identified were correct.

So we have 90 degree falls, 30 degree lifts, 5 weighted levers, 5 matching weights, 5 mechanisms, 10 pulleys, 5 lengths of cord and always 2 adjacent mechanisms working together. All we have to do is find a way to assemble them as Bessler did.

Thank you all for your comments so far, really very much appreciated. I’ve been doing this for too long on my own.

JC

[DrWhat]

Wonderful for you to share John.
I've been following from the sidelines.
The take home message from all this is present your ideas publicly early to minimise clinging to a design that has flaws, and to become too disheartened. I have been there many times.
You have a nice design that after keen eyes looking at it may get modified to become a runner. You never know.

[Fletcher]

I think most are sympathetic to your cause John, and tangibly want to help you realize it if they can contribute something worthwhile.

You've probably noted in previous forum sim builds over the years (Raj comes to mind as recent) that we usually build a sim of someones idea when it has flesh and detail on the bones to work with. Usually the person has settled on a combination that they think has the potential required and perhaps a mid-step is desired before the time and effort of a real world build, or as a cross-check on reality.

That's because as digital builders, predominantly of our own ideas, we can build and save to file tens if not hundreds of variations on a theme in days and weeks. Before whittling down possibilities and ending up at a conclusion about the particular empirical approach used and leaving no stone unturned. Usually the approach is abandoned before then when trends are observed on which future conclusions can be reliably predicted.

You said earlier that you are convinced you have the correct elements identified but there exists the possibility you might have interpreted them wrongly. I think there will be volunteers to sim your concept (and tweaks) when you settle on some connectivity combination to explore thoroughly.

I know I've gently pushed for this previously but I strongly suggest you procure a sim program and start practicing building your own sims and their iterations that inevitably lead from that investigation process. You will be amazed how quickly (in hours) you can start building wheels and changing things out etc. Even at 75 it is still very doable, and FUN ! The tutorials are helpful and educational. You'd be up and running in no time. And you always have the experience of this discussion board to call upon if needed.

I can only say from personal experience that I have made hundreds if not thousands of multiple lever-weight wheels with pulley connections etc. From the Driver located in and falling close into the CoR to minimize Cf's to too many to remember. It was one of the earlier and lengthy exercises that finally rammed home to me that 'height for width' was real and not theoretical. And that a workaround or mechanical cheat was required to beat it to have asymmetric torque conditions in a true gravity wheel. But I doubt I would have 100% believed it had I not done the hard yards with the lws, ropes and pulleys, springs, SBs etc, and thousands of sims (ok .. maybe not as many as KB). Finally the little corner of my small brain that said "what if" was now empty and free of ghosts. It now said "move on".

Best As Always.

-f

[John Collins]

Well I guess it’s never to late to learn a new skill and perhaps now is the time. Any advice on a suitable sim would be appreciated.

I might indeed have interpreted one or more clues wrongly and I’ve always been aware of the possibility, even the likelihood, but you do what you think is right and modify it as you go along, mental lightbulbs flashing on and off as inspirations come and go, which can be exhilarating.

Maybe I will manage to put up a sim of a creation of mine one day, but don’t hold your breath in anticipation, I expect I will try to find the fault in my thinking by manipulating by hand one of my new mechanical designs first.

Thank you sincerely Fletch and of course DrWhat and many more, for your kind words of encouragement

JC